Printer Friendly

Another look at epoxy thermosets correlating structure with mechanical properties.

INTRODUCTION

Epoxy prepolymers are classic building blocks for thermosetting composites because of the ease of fabrication, good flow, excellent mechanical and thermal performances including high resistance to solvents and corrosive environments, excellent fiber wetting, and low shrinkage upon curing [1-6]. Highly crosslinked matrix materials are practical for use in many applications as most of the starting materials have sufficiently low viscosities for processing, offering the best condition for blending and ease of handling in liquid form. Often, the glassy epoxy thermosetting polymers must be highly crosslinked to develop the highest possible modulus and glass transition temperature ([T.sub.g]). Conversely, the highly crosslinked network imparts inherent brittleness to glassy epoxy thermosets. The thermosetting matrixes have performance characteristics that depend on the network molecular architecture and connectivity, for example, less conversion results in a higher number of dangling ends, albeit temporary in many cases. The starting materials require high conversion to produce stable matrixes; however, as crosslink density (vc) increases, the network mobility and extent of crosslinking become interrelated. This leads to a performance limiting factor especially regarding toughness or force distribution over longer distances and shorter time scales as the building blocks become restricted in their mobility. Differential scanning calorimetry (DSC), infrared spectroscopy (IR), Raman spectroscopy, and fluorescence spectroscopy have been used to study the extent of epoxy reaction and its kinetics [3, 7-23]. Cure kinetics via DSC indicates the overall extent of chemical change but does not provide specifics regarding cure chemistry at molecular level. Aliphatic amines quench the fluorescence intensity of aromatic hydrocarbons and themselves do not fluoresce, hence intrinsic fluorescence characterization is inapplicable for epoxy prepolymers cured with aliphatic amines [22], Mid-FTIR (600-4000 [cm.sup.-1]) characterization has been used to measure epoxy functional group reaction kinetics; however, the methods are not capable of providing clear distinction between epoxy, primary, and secondary amine concentration changes due to band overlap. Conversely, near infrared (NIR) spectroscopy characterization can differentiate among distinct bands specific to each epoxide, primary amine, phenyl, combined primary-secondary amine, and the resulting hydroxyl groups during cure of epoxy-amine systems [12-19]. The cure kinetics and mechanisms of epoxy-amine reactions studied via NIR spectroscopy have been reported by some researchers [13, 15, 16, 24], Diaminodiphenylsulfone (DDS) when cured with epoxy prepolymers provides thermosets of high mechanical and thermal properties due to its rigid and aromatic structure [25, 26]. Cure kinetics and properties of DDS cured epoxy thermosets have been studied mostly with 4,4'-DDS version [13-16, 21, 24-35] while a very limited number of documents on 3,3'-DDS cured epoxies is available [36, 37],

Mechanical and thermal properties of 4,4'-DDS cured epoxy thermosets with diglycidyl ether of bisphenol-A (DGEBA), triglycidyl aminophenol (TGAP), and tetraglycidyl-4,4-diaminodiphenylmethane (TGDDM)based epoxy prepolymers have been quoted from literature (displayed in Table 1). [T.sub.g] ~ 245[degrees]C of epoxy thermoset with TGDDM and 3,3'-DDS [36] and notch toughness in the range 4-15 MPa x [m.sup.1/2] of DGEBA and 3,3'-DDS epoxy thermoset [37] have been reported.

Nonstoichiometric formulation of epoxy-crosslinker thermosetting systems results in varying network structure and properties. Stoichiometric formulation of an epoxy-amine system is expected to yield varying networks and properties at varying extent of epoxy-amine reaction. In case of 3,3'DDS and 4,4'-DDS (meta and para isomers of DDS, respectively), several studies on cure kinetics [21, 24], properties of modified [25, 26, 28, 29, 31-33, 35] and unmodified [27, 34], extent of reaction versus properties [30] of epoxy--4,4'-DDS systems have been reported. Although 3,3'-DDS cured epoxy thermosets were reported to have excellent matrix properties [36, 37] comparable with 4,4'-DDS based epoxy thermosets, there is a lack of data on structure-property relation as a function of epoxyamine reaction in 3,3'-DDS based epoxy thermosets. In this study, a DGEBA-based epoxy prepolymer of molecular weight ~380 g/mol was selected to synthesize epoxy thermosets of varying network characteristics at varying extent of reaction with 3,3'-DDS. The aim of this research focuses on characterizing varying network architectures of a stoichiometrically balanced blend of the DGEBA-based epoxy prepolymer and 3,3'-DDS. Varying degrees of conversion of the stoichiometric blend were developed using different thermal cure profiles and by allowing vitrification. This in turn resulted in different degrees of branching and linear growth as tracked via near IR cure kinetics. Mechanical and thermal properties were evaluated and correlated with the extent of cure, network architecture, and connectivity of the system. An empirical equation was also designed to relate chemical crosslinking with the extent of cure.

EXPERIMENTAL

Materials

A diglycidyl ether bisphenol-A (DGEBA)-based liquid epoxy prepolymer [reported epoxy equivalent weight (EEW) of 185-192 g/equiv., density 1.16 g/[cm.sup.3] at 25[degrees]C, and viscosity 110-150 poise at 25[degrees]C] was donated by Hexion (Momentive). Prior to use, the epoxide equivalent weight (EEW) of DGEBA-based epoxy prepolymer was determined via ASTM D1652-97 to be 188.72. 3,3'-diaminodiphenylsulfone (3,3'-DDS) with 98% pure was purchased from TCI, America [Amine hydrogen equivalent weight (AHEW) 63.26], Figure 1 displays basic chemical structure of DGEBA-based epoxy prepolymer and crosslinker. All the chemicals were used as received unless mentioned particularly.

Specimen Preparation and DSC Analysis

Epoxy prepolymer was placed in a vacuum oven at 100[degrees]C for 3 h to remove any volatile components and then cooled to ambient. Ten grams (0.0530 equiv.) of epoxy prepolymer was weighed into a scintillation vial and 3,3'-DDS (3.3523 g, 0.0530 equiv., adjusted for 98% purity) was added to the same vial. The vial contents were mixed for 1 min at 2700 rpm in a centrifugal Speed Mixer (FlackTek, Inc.). Approximately 7.5 mg of this stoichiometric blend was weighed into aluminum DSC pans and sealed for DSC analysis on a TA Instruments, DSC Q 2000. The results were processed via Universal Analysis 2000 software. To select temperature and determine energy barrier in combination with maximum rate for the epoxy-amine reaction in DGEBA-based epoxy prepolymer and 3,3'-DDS blend, DSC was performed at 1.0, 1.5, 2.0, and 2.5[degrees]C/min heating rate under nitrogen environment. Each experiment was conducted in duplicate and the average results were reported here.

Specimen Preparation, NIR Spectroscopy, and Cure Kinetics

The epoxy prepolymer, 10 g (0.0530 equiv.), was weighed into a scintillation vial, degassed, and then blended with 3.3523 g (0.0530 equiv.) of 3,3'-DDS in scintillation vial. The vial was placed in an oil bath maintained at 120-125[degrees]C and stirred mechanically for 30-40 min to ensure that 3,3'-DDS dissolved in the epoxy prepolymer. The hot blend was then sandwiched between two glass microscope slides using a microscope cover glass as a spacer to achieve a standardized sample thickness of 0.20 mm. Each sample was cured as described in Table 2. NIR spectra were obtained using a Nicolet 6700 FT-IR Analyzer from Thermo Scientific equipped with an InGaAs detector, [CaF.sub.2] beam splitter, and a quartz-halogen source. Absorption spectra were recorded in the NIR region of 4000-8000 [cm.sup.-1] (32 scans, resolution 4 [cm.sup.-1]). The epoxy, primary amine, and secondary amine concentrations were calculated using the integrated area under each respective peak using Lambert-Beer's law while the total amine H concentration was calculated by adding the primary amine H and secondary amine H concentrations from each spectrum. Lambert-Beer's law is normally expressed as [13]

A = [log.sub.10] ([I.sub.0]/I.sub.t]) [epsilon]bc, (1)

where A, [I.sub.0], [I.sub.t], [epsilon], b, and c are absorbance, intensity of incident radiation, intensity of transmitted radiation, molar extinction coefficient [L/(mol cm)], path length of radiation passed through the sample, and molar concentration (mol/L), respectively. The molar extinction coefficient is a unique property of each material/functional group at a particular frequency and is widely used for quantitative determination. By representing the concentration in terms of mol/kg, Eq. 1 can be simplified (for standard sample thickness or constant path length) as:

A = [alpha]c (2)

where [alpha] = [epsilon]b, and [alpha] is referred to as molar absorptivity (kg/mol) and c--molar concentration (mol/kg). As molar absorbance (A) is an additive property, Eq. 2 for n components can be written as:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (3)

where [SIGMA]A is total absorbance (i.e., combined band area) of components 1 to n.

SAMPLE PREPARATION FOR TESTING MATERIALS PROPERTIES

Degassed epoxy prepolymer (140 g) was added to 46.93 g 3,3'-DDS in a 250 mL beaker and placed in an oil bath at 125[degrees]C. The blend was stirred mechanically until the contents turned transparent. The hot blend was then degassed for ~3 min and cast into preheated (125[degrees]C) silicon molds to form bars with the appropriate dimensions for testing compression, flexural, tensile, DMA, and fracture toughness. Sample compositions and cure schedule are summarized in Table 2.

CHARACTERIZATION OF PHYSICAL PROPERTIES

Dynamic mechanical analysis was conducted on a DMA Q800 from TA Instruments, Inc. in tension mode using free films. The free films (~1-mm thick, ~5-mm wide) were characterized at a frequency of 1.0 Hz using a heating rate of 2[degrees]C/min with the DMA set for strain control. Each DMA dataset represents the average of three specimens for each formulation. The films were monitored for storage and loss moduli as well as mechanical [T.sub.g] from the tan [delta] maxima. The molecular weight between crosslinks ([M.sub.c]), defined as the sample weight in grams that contains 1 mole of elastically effective chains, was used to define each sample's momentary network connectivity. [M.sub.c] was determined from the storage modulus in the rubbery plateau region. Storage modulus (E') is inversely proportional to the chain length between entanglement molecular weight ([M.sub.c]) as [38-40]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (4)

where q is front factor (usually taken as 1.0 and followed for our calculations), p is density (g/[cm.sup.3]) and was acquired using TMA at [T.sub.g] +50[degrees]C, R is universal gas constant [R = 8.314 J/(mol K)] and [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] is the storage modulus (MPa) at [T.sub.g] + 50[degrees]C temperature. Crosslink density ([v.sub.c]) is the moles of elastically effective network chains in unit volume and was calculated by Eq. 5 as [41]:

[v.sub.c] = E'/3RT (5)

Density measurements for the samples based solely on epoxy were performed using Mettler Toledo analytical balances at room temperature. If [[rho].sub.0] is density of auxiliary liquid, the density ([rho]) of any solid can in principle be written as [42]:

[rho] = A / A - B ([[rho].sub.0] - [[rho].sub.L]) + [[rho].sub.L] (6)

where A is weight of sample in air, B is weight of the sample in the auxiliary liquid (water), [[rho].sub.L] is air density (0.0012 g/[cm.sup.3]). Sample density above the [T.sub.g] was determined using a thermal expansion coefficient on a Q400 Thermomechanical Analyzer operated at force of 0.02 N, heating rate 5 [degrees]C/min, and probe diameter 2.85 mm.

Fracture toughness properties were measured on a MTS Insight instrument. Fracture toughness samples were prepared with width 15.12 ([+ or -]0.03) mm, thickness 7.57 ([+ or -]0.08) mm, and crack length 7.15 ([+ or -]0.32) mm. The tests were operated in flexural mode with single-edge-notch bending (SENB) geometric arrangement according to ASTM D 5045-99, with a load cell of 10 kN (2248 Ibf) and the cross head speed of 1.27 mm/min. Fracture toughness, stress intensity factor ([K.sub.Ic]) was calculated using Eq. 7.

[K.sub.Ic] = ([P.sub.Q]/[BW.sup.1/2])f(x) (7)

where [P.sub.Q] is peak load (kN), B is thickness (cm), W is width (cm), and f(x) is calibration factor that depends on the ratio of crack length (a) and width, that is, x = a/W and f(x) was determined using Eq. 8

f(x) = 6[x.sup.1/2] [1.99 - x(1 - x) (2.15 - 3.93.y + 2.7[x.sup.2])]

[(1 + 2x).sup.-1][(1 - x).sup.-3/2] (8)

Plain strain condition was verified with [K.sub.Ic] using the following criterion:

B [greater than or equal to] 2.5 [([K.sub.Ic]/[[sigma].sub.y]).sup.2]

where [[sigma].sub.y] is yield stress. The averaged result of five specimens was reported.

Tensile properties were measured on a MTS Insight instrument equipped with a load cell of 10 kN (2248.09 lbf). Test was conducted according to the ASTM D 63803. Dumbbell-shaped tensile specimens were prepared with narrow section width 12.65 ([+ or -]0.02) mm, wide section width 19.0 ([+ or -]0.02) mm, thickness 3.08 ([+ or -]0.02) mm, narrow section length 60 mm, and overall length 165 mm.

Flexural tests were conducted on a MTS Insight with a 10 kN (2248.09 lbf) load cell and crosshead speed of 5 mm/min. Samples with thickness of 3.0 ([+ or -]0.15) mm, width of 13.0 ([+ or -]0.5) mm, and support span of 16 times the sample thickness were tested according to ASTM D 790.

Compression properties were measured on a MTS 810 Materials Test System (from MTS System Corporation) with appropriate sample clamps. Compression tests were conducted according to ASTM D 695-02a. Cylindrical samples were prepared with a length of 25.4 ([+ or -]0.01) mm and diameter of 12.7 ([+ or -]0.01) mm. The tests were conducted with a 111.2 kN (25,000 lbf) load cell at a crosshead speed of 1.27 mm/min without the aid of an extensometer. Modulus of elasticity was calculated from the slope of initial linear portion of the stress-strain curve. Compression analysis yielded modulus of elasticity, yield stress, and strain-at-yield.

Wide angle XRD of samples was recorded on a Rigaku Ultima3 X-ray diffractometer with a fine structure air-insulated X-ray tube (type FK 60-04) with a copper anode (Cu K[alpha]), scan speed of 0.5[degrees]/min and scan from 5[degrees] to 50[degrees].

RESULTS AND DISCUSSION

Epoxy-amine reaction/cure conditions of DGEBA epoxy prepolymer with 3,3'-DDS were determined via DSC analysis of reaction exotherm. Figure 2 shows DSC thermograms that were acquired using four distinct heating rates, 1.0, 1.5, 2.0, and 2.5[degrees]C/min under nitrogen environment to detennine the initial temperature ([T.sub.i])), peak temperature ([T.sub.p]), end temperature ([T.sub.f]), and heat of exotherm ([DELTA]H) during cure of DGEBA epoxy prepolymer and 3,3'-DDS (Table 3). The DGEBA-based epoxy prepolymer reacted initially with 3,3'-DDS at ~100-110[degrees]C and exhibited a maximum reaction rate between 155 and 178[degrees]C. [T.sub.i] [T.sub.P], and [T.sub.f] shifted to higher temperatures with increasing heating rates as was reported by other researchers [43], The maximum fluctuation in [DELTA]H values obtained across the four different heating rates was within 17.1 J/g. The energy barrier of the DGEBA-based epoxy prepolymer--3,3'-DDS reaction was calculated using the Kissinger and the Ozawa-Flynn-Wall models. The Kissinger model is expressed as [43, 44]:

-ln ([beta]/[T.sub.P.sup.2]) = ln(AR/[E.sub.K]) + [E.sub.K]/[RT.sub.P] (10)

where [beta] is heating rate, A is pre-exponential factor, R is universal gas constant, [E.sub.K] is activation energy of the reaction, and [T.sub.P] is peak temperature at which the reaction rate is maximum. The plot of -ln([beta][T.sub.P.sup.2] versus 1/[T.sub.P] results in a straight line and the activation energy, [E.sub.K], can be calculated from its slope. The Ozawa-Flynn-Wall model based on Doyle's approximation is expressed as [44]:

log [beta] = log{[AE.sub.O]/g([alpha])R} - c - l[E.sub.O]/RT (11)

where [E.sub.O] is activation energy of the reaction, g([alpha]) is conversion dependent function, and c and l are couple tabulated coefficients. Commonly, c = 2.313 and l = 0.4567 if [E.sub.O]/RT = 28-50 and c = 2.000, l = 0.4667 if [E.sub.O]/RT = 1830, and c= 1.600 and l = 0.4880 if [E.sub.O]/RT = 13-20. In most aromatic diamine and epoxy prepolymer reactions, [E.sub.O]/RT = 13-20. We use c = 1.600 and l = 0.4880 for our calculation. Specific to the co-reaction of an aromatic diamine and DGEBA epoxy, the Ozawa-Flynn-Wall equation was therefore rewritten as [44]:

log [beta] = log {[AE.sub.O]/g([alpha])R} - 1.600 - 0.4880 [E.sub.O]/[RT.sub.p] (12)

The graph of log[beta] versus I/[T.sub.P] yields a straight line. Activation energy ([E.sub.O]) was determined from the slope of this straight line. The energy of activation for this epoxy-amine reaction was determined via both the Kissinger and Ozawa-Flynn-Wall methods (shown in Figs. 3 and 4, respectively). The resulting activation energy values calculated by Kissinger and Ozawa-Flynn-Wall methods were very close, that is, 60.93 kJ/mol and 60.73 kJ/mol. The commonly reported 61-69 kJ/mol activation energy for epoxy-amine reactions specific to DGEBA epoxy with aromatic diamine provided additional support for our experimental results [11, 24], Further analysis was performed using the Friedman equation to assess the activation energy across conversion associated with the reaction using an isoconversion method [44],

ln (d[alpha]/dt) = ln {f(alpha])A} - [E.sub.a]/RT (13)

where [E.sub.a] is activation energy and f([alpha]) is unknown reaction model. [E.sub.a] can be calculated from the slope of the straight line from the plot of ln(d[alpha]/dt) versus l/T. The activation energy versus conversion calculated via the Friedman method is displayed in Fig. 5. It is worthy of note that the activation energy during cure initially decreased from ~62.00 to ~55.00 kJ/mol up to 60% conversion ([alpha] = 0.10-0.60) and then increased to ~60 kJ/mol at higher degrees of conversion. The temporarily decreased activation energy level was attributed to the autocatalytic effect driven by the formation of tertiary amine once the secondary amines reacted with the epoxy prepolymer. The upward trend in [E.sub.a] at higher conversion was attributed to higher viscosity whereby diffusion restricted mobility for active polymer chains begins to impact the [E.sub.a] term as reported elsewhere [44].

NIR spectra of 3,3'-DDS and DGEBA-based epoxy prepolymer are displayed in Fig. 6. 3,3'-DDS exhibited characteristic bands at 6696 [cm.sup.-1] for overtone of--N[H.sub.2], 6077-5957 [cm.sup.-1] for overtone of aromatic C--H, 5061-4769 [cm.sup.-1] for combination mode of stretching and bending of--N[H.sub.2], 4665-4641 [cm.sup.-1] for combination of aromatic C=C and C--H, and 4598-4535 [cm.sup.-1] for combination of C--N stretching and bending of--N[H.sub.2] [13]. In the NIR spectrum of DGEBA-based epoxy prepolymer, weak absorption at 7001 [cm.sup.-1] due to first overtone of--OH stretching ([approximately equal to]2 x 3507) was observed [13]. The bands detected at 6071 and 4530 cm 1 are attributed to the first overtone of terminal--C[H.sub.2]-- stretching ([approximately equal to]2 x 3056) and combination between the epoxy--C[H.sub.2]--stretching and the epoxy--C[H.sub.2]-- deformation ([approximately equal to] 3056 + 1456 [cm.sup.-1]), respectively [11, 13, 45], The bands observed at 5989-5656 [cm.sup.-1] are due to overtones of aromatic C--H ([approximately equal to]2 x 3037) and aliphatic C--H ([approximately equal to]2 x 2968, 2 x 2931, 2 x 2874, 2 x 2834) stretching [13, 46], The two bands at 4681 and 4622 [cm.sup.-1] noted from the combination of the aromatic C--H and C=C stretching ([approximately equal to]3037 + 1607, [approximately equal to]3037 + 1509) [13, 47],

The cure specific NIR spectra collected for DGEBA-based epoxy prepolymer with stoichiometric proportion of 3,3'-DDS at 90[degrees]C for 18 h and 150[degrees]C for 20 h is displayed in Fig. 7 (only the specific bands of interest have been highlighted). Each spectrum was normalized against an internal control peak, the phenyl band at 4622 [cm.sup.-1] (concentration remains unchanged during cure). The primary amine band at 5066 [cm.sup.-1], combined band of epoxy and primary amine groups at 4531 [cm.sup.-1], and a combined band of primary and secondary amine groups at 6569-6678 [cm.sup.-1] consistently progressed to lower intensity over time, whereas the hydroxyl band at 6991 [cm.sup.-1] increased in intensity as a result of the epoxyamine reaction. Similar results were observed when characterizing other cure profiles albeit only selected spectra are reported herein.

The primary amine concentration in mixture was determined from the band area of the band at 5060 [cm.sup.-1] as:

[A.sub.1] = [[alpha].sub.pa1] [c.sub.pa1] (14)

where [A.sub.1] is N[H.sub.2] band area at 5060 [cm.sup.-1], [c.sub.pa1] is N[H.sub.2] concentration (mol/kg) and [[alpha].sub.pa1] is molar absorptivity of N[H.sub.2] band at 5060 [cm.sup.-1] that was quantified as 1.9213 kg/mol from the control 3,3'-DDS monomer spectrum at a sample thickness 0.20 mm. The blended epoxy concentration was calculated using a combined band for the epoxy and primary amine at 4530 [cm.sup.-1] using the simplifying Eq. 3

[SIGMA][A.sub.2] = [[alpha].sub.e][c.sub.e] + [[alpha].sub.pa 2] [C.sub.pa 2] (15)

where [SIGMA][A.sub.2] is combined band area of epoxy and primary amine groups, [[alpha].sub.e] and [c.sub.e] are molar absorptivity and concentration of epoxy group; oepa2 and [c.sub.pa2] are molar absorptivity and concentration of primary amine group. [[alpha].sub.e] was calculated to be 0.6750 kg/mol from the spectrum of the DGEBA-based epoxy prepolymer at 0.20 mm thickness. Similarly, [[alpha].sub.pa2] was determined to be 0.4282 kg/mol from the 3,3'-DDS monomer spectrum at 0.20 mm thickness. [c.sub.pa2] is actually [c.sub.pa1] for the same spectrum. Secondary amine concentration in the mixture was determined from the band at 6696 [cm.sup.-1] after rearranging Eq. 3 as:

[SIGMA][A.sub.3] = [[alpha].sub.pa3] [C.sub.pa3] + [[alpha].sub.sa] [c.sub.sa] (16)

where [SIGMA][A.sub.3] is combined band area of primary amine and secondary groups, [[alpha].sub.pa3] and [c.sub.pa3] are molar absorptivity and concentration of primary amine group; and [[alpha].sub.sa] and [c.sub.sa] are molar absorptivity and concentration of secondary amine group. Assuming secondary amine concentration is zero just after mixing and [[alpha].sub.pa3] was determined from Eq. 16 when [c.sub.pa3] = [c.sub.pa1] for the same spectrum. Once, [[alpha].sub.pa3] and [c.sub.pa3] were known, [[alpha].sub.sa] was calculated assuming that the secondary amine does not react during the initial experiments. [[alpha].sub.pa3] and [[alpha].sub.sa] values were determined imposing the above conditions to be 1.6124 and 1.0306 kg/mol, respectively. The total amine hydrogen concentration in mixture was calculated as:

[c.sub.ta] [c.sub.pat] + [c.sub.sat] (17)

where [c.sub.pat], [c.sub.sat], and [c.sub.ta] are primary, secondary, and total amine hydrogen concentrations, respectively, at time t.

Figures 8-10 summarize calculated concentration and conversion results for the stoichiometric blend of epoxy and amine heated to 90[degrees]C and 150[degrees]C, 125[degrees]C and 200[degrees]C, and 175[degrees]C versus time, respectively. Within the figure, the -1 h data points represent the preblended initial concentration(s). Results confirmed that the primary amine concentration diminished rapidly during cure at 90[degrees]C and approached zero within 10 h while the secondary amine concentration increased initially to a maximum and exhibited slower changes after this point in time. The data sup' ported that the combination of a physical state transition and time-temperature relationship for functional group conversion, the reduced mobility, and the lower reactivity for the secondary amine and epoxy functionality of reaction product established a maximum concentration of secondary amine shortly after 6 h at 90[degrees]C. Afterward, the steadily decreasing concentration of secondary amine up to the 14 h data point at which the rate of decreasing concentration slowed dramatically up to the 18 h data point and was attributed to a vitrification and diffusion restricted state (during the timescale of analysis). The same material upon further heating up to 150[degrees]C exhibited a sharp NH concentration decrease over the additional 3 h as the remaining concentration approached the detectable limits (practically zero). It was observed that epoxy concentration and total amine hydrogen concentration decreased at almost an identical rate during cure at 90[degrees]C as well as at 150[degrees]C. The results indicated that the predominant concentration of primary amines was consumed via reaction with the epoxy (during cure at 90[degrees]C). Figure 8A shows a theoretical curve of secondary amine concentration, that is, secondary amine formed in situ did not react with epoxide, is shown. It is seen that the experimental curve of secondary amine concentration does not follow the theoretical curve. This further supported the slow but important consumption of secondary amine occurs even at vitrification-driven mobility restricted state when the [T.sub.g] of system was about 100.4[degrees]C which was higher than the cure temperature (90[degrees]C) (Fig. 11 and Table 4). The results revealed that the secondary amine does react with epoxy at 90[degrees]C, which in turn confirmed that the progression toward a glassy matrix occurs with linear but also branched growth patterns, that is, the secondary amine was not frozen from occurring. It was intended to force maximum linear growth before final cure to notice the contrast from matrix grown in a much more splayed manner. The functional groups, epoxy and total amine hydrogen conversions progressed steadily resulting in 75.67% and 75.73%, respectively, after 14 h at 90[degrees]C. Beyond 76% conversion, the functional group peaks diminish very slowly until the matrix was heated above 90[degrees]C. The semi-glassy matrix at 90[degrees]C reacted further but only marginally leading to epoxy and total amine conversions were calculated as 77.98% and 77.97%, respectively, at 18 h when the material [T.sub.g] was 113.04[degrees]C. Shifting the cure temperature to 150[degrees]C resulted in 96.04% epoxy conversion after 2 h where [T.sub.g] was 162[degrees]C. Similarly with vitrification limits reached during cure at the 90[degrees]C, extending the 150[degrees]C cure for an additional 18 h led to an epoxy conversion of 97.68% offering the material of [T.sub.g] 173.61[degrees]C.

The same stoichiometric material blend heated at 125[degrees]C exhibited a rapid decrease in primary amine concentration within the first hour (greater than 80% conversion) and was accompanied by a concurrent increase in the secondary amine concentration that reached a maximum concentration near the end of the first hour at 125[degrees]C. As noted earlier, the increasing viscosity as a glassy material approaches vitrification restricts the mobility of epoxy and amine groups and severely limits their conversion. Following the epoxy conversion at 4 h at 125[degrees]C, the system had converted 85.24% where [T.sub.g] was about 140.69[degrees]C. The attained higher viscosity concurrent with the reaction progress slowed conversion via vitrification sufficiently to limit the reaction progress to 86.62% epoxy conversion after 5 h at 125[degrees]C when the [T.sub.g] was 143.35[degrees]C. Using a post-cure protocol maintained at 200[degrees]C for 1 h resulted in an epoxy functional conversion of 99.47% and [T.sub.g] of 175.24[degrees]C. Furthermore, extending the cure process for an additional 4 h at 200[degrees]C resulted in the same degree of conversion (thought to be the practical limit). Considering the amine hydrogen conversion rate, it was observed that both conversion curves are almost identical throughout the curing process at 125[degrees]C and post-curing at 200[degrees]C. The data verified that amine and epoxy conversion were the result of mutual epoxy-amine reaction as was confirmed for the cure process using 90[degrees]C and post-curing at 150[degrees]C.

Figure 10 summarizes the time-specific concentration and conversion for stoichiometric epoxy amine reactions using 175[degrees]C as the reaction temperature. The reaction profile at 175[degrees]C resulted in epoxy and amine conversion during the first hour of 95.03% and 95.81%, respectively, where [T.sub.g] was 155.21[degrees]C. An additional 4 h heating resulted in a glassy material of 99.32% epoxy conversion and 99.37% amine conversion with [T.sub.g] of 175.48[degrees]C. Curing at 175[degrees]C also confirmed that the amines and epoxy reactions continue to be consumed concurrently, as observed with the previous cure profiles. To summarize, epoxy functional groups are mostly reacted with primary amine during the initial reaction, regardless of temperature. The primary amines are almost completely consumed in 10 h, 3 h, and 0.5 h at 90[degrees]C, 125[degrees]C, and 175[degrees]C, respectively. It was expected to observe several side reactions as noted in the literature, in particular, the etherification and homopolymerization reactions at cure profiles extended beyond epoxy conversion of 40-50% [8]. Other researchers have also reported that negligible etherification reactions occurred in stoichiometric blends of epoxy and aromatic amines when treated thermally at or below 200[degrees]C [48]. The reported activation energy, [E.sub.a], for each primary and secondary aromatic amine reaction with epoxy prepolymer ranges between 55 and 72 kJ/mol [8] while etherification and homopolymerization reactions have been shown to require 97-104 kJ/mol [8, 30] and 170 kJ/mol [30], respectively. [E.sub.a] values for the DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometric system ranged between 55 and 63 kJ/mol throughout the entire conversion process as measured from the DSC kinetics supporting that only epoxy-amine reaction occurred during curing of DGEBA epoxy prepolymer with stoichiometric amount of 3,3'-DDS. NIR spectral characterization further supported a DGEBA epoxy prepolymer reacting mostly with primary and secondary amines under selected cure conditions.

Critical extent of reaction at the gel point of bifunctional epoxy and tetrafunctional crosslinker is reported to be about 57.73% [49]. [T.sub.g] at any extent of reaction can be calculated using Eq. 18 if [T.sub.g0] and [T.sub.g[infinity]] of the epoxy amine blend are known [50]

[T.sub.g] - [T.sub.g0] / [T.sub.g[infinity]] - [T.sub.g0] = [lambda]x / 1 - (1 - [lambda])x (18)

where [T.sub.g0] and [T.sub.g[infinity]] are [T.sub.g]s at zero extent of reaction and complete reaction, respectively, while x and [lambda] are extent of reaction and adjustable parameter, respectively. [lamda] varies between 0 and 1. The value of [lambda] = 0.487 has been reported to be a good fit for bifunctional DGEBA-based epoxy resin system [50] and was also used in our study to develop [T.sub.g] versus fractional conversion (x) plot. [T.sub.g0] and [T.sub.g[infinity]] of DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometric blend were determined via DSC to be -12.86[degrees]C and 175.48[degrees]C, respectively. Figure 12 shows theoretical Tg as a function of fractional conversion plot of DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometric blend based on Eq. 18. It was noticed that experimental [T.sub.gS] are in good agreement with the theoretical ones. This means that [T.sub.gs] of DGEBA-based epoxy prepolymer and 3,3'-DDS can be well predicted as a function of epoxy conversion using the Eq. 18. Thus, [T.sub.g] at 57.73% epoxy conversion is expected to be about 62.4[degrees]C of DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometric blend. As initial curing temperatures were 90[degrees]C, 125[degrees]C, and 175[degrees]C which were above the [T.sub.g] at critical extent of reaction at the gel point, epoxy-amine reaction proceeded above the gel point conversion, owing to which the system gets vitrified well above the gel point conversion.

First-order kinetic in isothermal cure condition of epoxy-amine system is usually expressed by the following equation [11, 48]

ln ([C.sub.t]/[C.sub.0] = kt (19)

where [C.sub.t], and [C.sub.0] are initial concentration and concentration at time t, respectively, and k is rate constant. The slope of the straight line from the plot of ln([C.sub.t]/[C.sub.0]) versus t yields the value of k. Rate constant, k is related in Arrhenius equation as [48]:

In k = ln A - [E.sub.a]/RT (20)

where A, [E.sub.a], R, and T are pre-exponential factor, activation energy, universal gas constant, and temperature in Kelvin, respectively. Plot of Ink versus 1/T provides a straight line with a slope of -[E.sub.a]/R from which the activation energy can be calculated. Concentration of either epoxy or amine hydrogen at time t ([C.sub.t]) was calculated from NIR spectra of stoichiometric blend of DGEBA-based epoxy prepolymer and 3,3'-DDS during curing isothermally at 90[degrees]C, 125[degrees]C, and 175[degrees]C. Plots of In([C.sub.t]/[C.sub.0]) versus t for epoxy and amine hydrogen are shown in Fig. 13. We observed that the rate of decrease of ln([C.sub.t]/[C.sub.0]) for both the epoxy and amine hydrogen was higher when the blend was thermally maintained at higher temperatures reconfirming the higher rate of epoxy-amine reaction at 175[degrees]C. As the reaction proceeds, the decreasing rate of ln([C.sub.t]/[C.sub.0]) confirms the effect of reaching a diffusion controlled material state.

Figure 14 displays the lnk versus 1/T plot for both the epoxy and total amine hydrogen as extracted from their ln([C.sub.t]/[C.sub.0]) versus t plot during isothermal curing at 90[degrees]C, 125[degrees]C, and 175[degrees]C of DGEBA-based epoxy prepolymer and 3,3'-DDS. From the slope of the plot, activation energies of 53.39 and 53.18 kJ/mol were calculated for epoxy and total amine hydrogen, respectively. These values were approximately 13% lower than the activation energy determined via dynamic DSC kinetic analysis. DSC sample was prepared by mixing DGEBA-based epoxy prepolymer and 3,3'-DDS at ambient while NIR specimen was prepared by solubilizing 3,3'-DDS in DGEBA-based epoxy prepolymer at 125[degrees]C. Mixing at ambient yields no reaction between epoxy and amine whereas solubilizing 3,3'-DDS in DGEBA-based epoxy prepolymer at 125[degrees]C leads to an epoxy conversion of about 4-6% as noticed by NIR kinetic analysis. Activation energy ~36.73 kJ/ mol, specific to a DGEBA epoxy-aromatic amine reaction quantified via IR kinetic analysis has been reported [11]. The activation energy values of the epoxy-aromatic amine reaction as calculated from individual concentration changes of the epoxy and the amine over time match closely. This means that the dominated reaction under these conditions with these materials was epoxy with amine (primary and secondary).

DMA studies provided storage modulus and tan [delta] plots of DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometric systems cured at various cure schedules (shown in Fig. 15). Storage modulus (E') at ambient and [T.sub.g] (determined from tan [delta] peak) have been summarized in Table 5. Molecular weights between crosslinks ([M.sub.c]) calculated using Eq. 4 and crosslink density ([v.sub.c]) calculated using Eq. 5 for the stoichiometric blend of DGEBA-based epoxy prepolymer and 3,3'-DDS are also displayed in Table 5. The [T.sub.g]s of sample SI (128[degrees]C) and S3 (146[degrees]C) were lower than that of samples S2 (177[degrees]C), S5 (183[degrees]C), and S4 (184[degrees]C). Lower epoxy conversions in SI and S3 resulted in lower crosslink density, higher crosslink molecular weights, and lower [T.sub.g]. Samples S4 and S5 exhibited higher [T.sub.g]s within a fluctuation of one unit because of higher crosslink density resulting from similar extent of epoxy-amine reaction. It was observed that glassy E' decreases with increase in epoxy conversion while rubbery E' has reverse trend. The slopes of the rubbery E' curve of SI and S3 were greater than that of S2, S4, and S5 because of additional cure during heating in DMA. Cohesive energy decreases as a result of reduction in nonbonded intermolecular interactions as epoxy cure progresses [30]. Higher cohesive energy, that is, intermolecular interactions in sample SI and S3 resulted in higher modulus in the glassy state. In the rubbery state, storage moduli of S2, S4, and S5 were higher than that of SI and S3 due to their higher crosslink density. These results indicate that the modulus in glassy state is influenced by intermolecular interaction forces while rubbery state modulus is affected by crosslink density/chemical connectivity and network structure variability.

Density at [T.sub.g] + 50[degrees]C required to calculate [M.sub.c] is shown in Fig. 16. Ambient density of all samples was expectedly higher than the density at [T.sub.g] + 50[degrees]C. Density at ambient of samples S1-S5 did not fluctuate with epoxy conversion while density at [T.sub.g] + 50[degrees]C decreased with temperature a bit from 77% to 97% epoxy conversion. Extent of epoxy conversion of S1-S5 samples was in the order of S1 < S3 < S2 < S5 [approximately equal to] S4 as determined via near IR spectroscopy. [T.sub.g] + 50[degrees]C temperature of S1-S5 samples followed the increasing order of extent of epoxy conversion. Material density decreases with increase in temperature. The density of samples SI [right arrow] S3 [right arrow] S2 at [T.sub.g] + 50[degrees]C decreased as expected with increasing temperature [T.sub.g] + 50[degrees]C (178 [right arrow] 196 [right arrow] 227[degrees]C). Densities of S5 and S4 samples at [T.sub.g] -I- 50[degrees]C were almost equal but little higher than that of S2. This small increase in density at [T.sub.g] + 50[degrees]C from S2 [right arrow] S5 or S5 indicated a tighter network due to greater extent of epoxy reaction although [T.sub.g] + 50[degrees]C temperature of S4 is only 7[degrees] higher compared to S2.

Average value of molecular weight between crosslinks ([M.sub.c]) for fully converted epoxy-amine system having multifunctional crosslinker and bifunctional chain extender can be calculated as [51]:

[M.sub.c] = 2/[f.sub.cav] [M.sub.pc] (21)

where [M.sub.pc] is average molecular weight per crosslink, [f.sub.cav] is average crosslink functionality. [M.sub.pc] and [f.sub.cav] can be calculated as [51]:

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (22)

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (23)

where [M.sub.E] is EEW of epoxy prepolymer, [M.sub.f] is molecular weight of fth functional amine, f is the amine functionality, [[PHI].sub.f] is mole fraction of amine hydrogen provided by the /th functional amine. Application of Eqs. 21-23 in DGEBA-based epoxy prepolymer and 3,3'-DDS stoichiometrically balanced system yielded [M.sub.c] = 501.44, [M.sub.pc] = (4[M.sub.e] + [M.sub.f]) = 1002.88, and [f.sub.cav] = 4, respectively. This means that the crosslinkable unit mass is a DGEBADDS-DGEBA trimer. Depending on the number of chemical crosslinks ([n.sub.c]) in a crosslinkable unit mass, the average chemical crosslink molecular weight ([M.sub.c]) can be determined with the following empirical equation as:

[M.sub.c] = (2[M.sub.E] + [M.sub.A]/[n.sub.c], (24)

where [M.sub.E] and [M.sub.A] are molecular weights of bifunctional epoxy prepolymer and tetrafunctional crosslinker, respectively. nc is the number of chemical crosslinks ([n.sub.c] > 1). [n.sub.c] of stoichiometrically balanced bifunctional epoxy and tetrafunctional crosslinker system can be calculated using Eq. 25 as:

[n.sub.c] = 1 + (4x - 3) (25)

where x is extent of reaction (1 [greater than or equal to] x [greater than or equal to] 0.75). 0.75 is the critical extent of reaction for stoichiometrically balanced bifunctional epoxy and tetrafunctional crosslinker system.

According to Carothers, the critical extent of reaction ([p.sub.c]) and average functionality ([f.sub.avg]) of the reactants are related as [49]:

[p.sub.c] = 2/[f.sub.avg] (26)

Average functionality ([f.sub.avg]) of stoichiometrically balanced bifunctional and tetrafunctional reactive monomers system was calculated to be 2.67 using the following equation [49]

[f.sub.avg] = [SIGMA][N.sub.i][f.sub.i]/[SIGMA][N.sub.i] (27)

where [N.sub.i] is the number of molecules of monomer i with functionality [f.sub.i]. Figure 17 displays theoretically predicted Mc (calculated using Eqs. 24 and 25) values as a function of epoxy conversion ([greater than or equal to]0.75) as well as experimentally determined [M.sub.c] values for samples S1-S5. It was observed that experimentally determined [M.sub.c] values for samples S2-S5 were very close to the [M.sub.c] values determined via empirical Eq. 24. Thus, Eq. 24 can be used to predict [M.sub.c] of an epoxy-amine system across its conversion and beyond the critical extent of reaction. This equation does not encompass physical crosslinks that occur during highly crosslinked and high conversion situations. [M.sub.c] value of SI (epoxy conversion ~78%) was about 200 units higher than the value predicted by Eq. 24. We present a very simplified sketch of the probable chemically connected network at 50%, 75%, and 100% conversions of a stoichiometrically balanced DGEBA-based epoxy prepolymer and 3,3'-DDS systems in Fig. 18. The cartoon was meant to aid an understanding how [M.sub.c] values could be higher than the predicted value. If the epoxy-amine reaction from DGEBA-based epoxy prepolymer and 3,3'DDS proceeds as shown in Fig. 18, there is a possibility of network formation with higher linear chain length than trimer (DGEBA-DDS-DGEBA). The systems could possess very few chemical crosslinks even at epoxy conversion close to 75%, leading to the higher than predicted [M.sub.c] value from Eq. 24. Experimental [M.sub.c] values lower than the expected values for fully converted stoichiometrically balanced thermoset systems are most likely due to physical crosslinks, however, we have not proven this to be definitive with our research alone.

Fracture toughness results ([K.sub.Ic]) are reported as a function of the cure profiles in Fig. 19. [K.sub.Ic] values decreased consistently with increasing epoxy conversion. Samples SI and S3 had lower crosslink density because of lower epoxy conversion beyond the critical extent of reaction resulting in higher fracture toughness. In addition, the higher concentration of pendant chain ends from the unreacted epoxy oligomer resulted in higher flexibility compared to fully cured samples. The same samples, that is, under cured materials consistently provided broader tan delta peaks as measured by the full width at half height (FWHH) values (Table 5). The results suggest that dangling ends from the unreacted epoxy fractions and lower crosslink density provided higher fracture toughness from the partially cured epoxy-amine thermoset beyond the critical extent of conversion. A reverse trend, that is, Kic increased with conversion beyond gel point has been reported for epoxy thermoset formed from DGEBA-based epoxy prepolymer and 4,4'-DDS [30].

Flexural, tensile, and compression properties of partially and fully cured epoxy-amine thermosets from DGEBA based epoxy prepolymer and 3,3'-DDS are displayed in Fig. 20. It was observed that flexural strength increased from 101 MPa to 141 MPa with increase in epoxy conversion from 78% to 87% epoxy conversion and beyond 87% epoxy conversion there was no significant increase in flexural strength. The flexural strength values between 87% and 99% epoxy conversion were within the standard deviation limit. The flexural modulus values decreased from 3.94 GPa to 2.95 GPa with respect to epoxy conversion from 78% to 99%. Tensile and compression moduli followed a similar trend. Tensile strength increased from ~70 MPa to ~93 MPa with increasing epoxy conversion from ~78% to 99% in a similar fashion as noticed in flexural strength. Compression yield stresses for all cure schedules were within the range of 137-148 MPa. Tensile strain-at-break, flexural strain-at-peak stress (SI sample only broke at 2.63%), and compression yield stress increased with increase in epoxy conversion. Higher flexural, tensile, and compression moduli at lower conversion were found to be dominated by nonbonded intermolecular forces over chemical connection in glassy state [30]. It appeared that that the higher noncovalent intermolecular attractive forces drive the increased cohesive energy and result in increased resistance to deformation. Liu et al. reported that the ring opening of both epoxy groups of DGEBA-based epoxy prepolymer forming secondary hydroxyls increases cohesive energy by about 13.4 kJ/mol while 4,4'-DDS decreases cohesive energy by about 27 kJ/mol when forming tertiary amine moieties during cure [52]. This means that the curing of DGEBA-based epoxy prepolymer with DDS decreases the cohesive energy overall as the monomer blend proceeds to form a cured thermoset. Conversely, tensile and flexural strength are probably ruled by chemical connectivity due to the greater strain and mobilized network drawn upon during these deformation modes. That is the reason why tensile and flexural strength increase with increasing epoxy conversion.

In epoxy thermosets, cooling below the [T.sub.g] often brings about physical aging. Endotherms were observed in DSC thermograms of undercured samples (S1-S3) probably due to aging. Peak of these endotherms were taken as [T.sub.g] of the materials. To understand this system further, the same materials were analyzed via XRD to check for some sort of ordering (although very small) in the network, which could contribute to material properties at lower epoxy conversion. 3,3'-DDS is a crystalline solid material with a melting temperature ~170[degrees]C. A sharp crystalline peak at an angle 2[theta] = 44[degrees]-45[degrees] was observed in XRD curve of SI, S2, S3, and S5 samples as shown in Fig. 21 indicating some sort of ordering in the cured network.

CONCLUSION

DGEBA-based epoxy prepolymer and 3,3'-DDS were blended at stoichiometric proportions and cured under various cure schedules to obtained different networks with varying extent of epoxy-amine reaction to observe the effect of network formation on performance properties of cured thermosets. The energy of activation for epoxyamine reaction of DGEBA-based epoxy prepolymer and 3,3'-DDS was about 61 kJ/mol as calculated by DSC. Curing at 90[degrees]C and 125[degrees]C led to maximum vitrification limited epoxy conversion of 78% and 87%, respectively. The cure profiles validated the necessity of the post-curing to have optimum performance properties. Curing of DGEBA-based epoxy prepolymer with 3,3'-DDS at 90[degrees]C was targeted to have as much as possible linear growth by facilitating primary amine reactions selective over secondary amine reactions, yet the different cure profiles ultimately resulting in crosslinking as measured by DMA. NIR kinetic analysis confirmed that the epoxy-amine reaction occurred predominantly even at cure temperatures above 125[degrees]C (from a stoichiometric blend of DGEBA-based epoxy with 3,3'-DDS). Higher storage, tensile, flexural, and compression moduli at lower epoxy conversion of DGEBA-based epoxy prepolymer and 3,3'-DDS thermoset were observed due to the presence of higher intermolecular interaction forces, confirming what was noticed by other researchers [30, 52]. Higher strength at higher epoxy conversion was controlled by high extent of crosslinks. Dangling ends/ pendant functional groups and unreacted epoxy oligomer fractions impart flexibility, lower crosslink density and higher fracture toughness at 78-87% epoxy conversions as the network has more degrees of freedom for distribution of deformation forces without chain scission. Additionally, results suggested that maintaining a small excess of epoxy over the stoichiometric amount of amine would offer improved performance properties. Experimental [M.sub.c] values were compared with theoretically calculated by empirical Eqs. 24 and 25 that were developed by combination of established equations (Eqs. 21-23, 26, 27). It was turned out to be a valid equation which can be used to predict molecular weight between crosslinks ([M.sub.c]) with respect to epoxy conversion.

ACKNOWLEDGMENT

The authors acknowledge Northrop Grumman Shipbuilding, Inc., Gulfport, Mississippi for their material donation and support.

REFERENCES

[1.] G. Pan, Z. Du, C. Zhang, C. Li, X. Yang, and H. Li, Polymer, 48, 3686 (2007).

[2.] P. Musto, E. Martuscelli, G. Ragosta, P. Russo, and P. Villano, J. Appl. Polym. Sci., 74, 532 (1999).

[3.] P. Musto, M. Abbate, G. Ragosta, and G. Scarinzi, Polymer, 48, 3703 (2007).

[4.] C. Li, G. A. Medvedev, E. W. Lee, J. Kim, J. M. Caruthers, and A. Strachan A, Polymer, 53, 4222 (2012).

[5.] J. Chen and A.C. Taylor, J. Mater. Sci., 47, 4546 (2012).

[6.] T.H. Hsieh, A.J. Kinloch, A.C. Taylor, and I.A. Kinloch, J. Mater. Sci., 46, 7525 (2011).

[7.] M. Ghaemy, S.M.A. Nasab, and M. Barghamadi, J. Appl. Polym. Sci., 104, 3855 (2007).

[8.] N. Sbirrazzuoli, A. Mititelu-Mija, L. Vincent, and C. Alzina, Thermochim. Acta, 447, 167 (2006).

[9.] P.P. Adroja, R.Y. Ghumara, and P.H. Parsania, J. Appl. Polym. Sci., 130, 572 (2013).

[10.] S.G. Hong and C.S. Wu, Thermochim. Acta, 316, 167 (1998).

[11.] M. Ghaemy, A.H. Shahriari, and S.M.A. Nasab, Iranian Polym. J., 16, 829 (2007).

[12.] R.J. Varley, G.R. Heath, D.G. Hawthorne, J.H. Hodgkin, and G.P. Simon, Polymer, 36, 1347 (1995).

[13.] B.G. Min, Z.H. Stachurski, J.H. Hodgkin, and G.R. Heath, Polymer, 34, 3620 (1993).

[14.] B.G. Min, Z.H. Stachurski, and J.H. Hodgkin, Polymer, 34, 4908 (1993).

[15.] N.A. St. John and G.A. George, Polymer, 33, 2679 (1992).

[16.] C.J. de Bakker, N.A. St. John, and G.A. George, Polymer, 34, 716 (1993).

[17.] L. Xu, J.H. Fu, and J.R. Schlup, Ind. Eng. Client. Res., 35, 963 (1996).

[18.] C. Billaud, M. Vandeuren, R. Legras, and V. Carlier, Appl. Spectros., 56, 1413 (2002).

[19.] G. Ragosta, P. Musto, M. Abbate, and G. Scarinzi, Polymer, 50, 5518 (2009).

[20.] R. Hardis, J.L.P. Jessop, F.E. Peters, and M.R. Kessler, Compos. A, 49, 100 (2013).

[21.] J.C. Song and C.S.P. Sung, Macromolecules, 26, 4818 (1993).

[22.] A. Rigail-Cedeno and C.S.P. Sung, Polymer, 46, 9378 (2005).

[23.] E. Pyun and C.S.P. Sung, Macromolecules, 24, 855 (1991).

[24.] D.J.T. Hill, G.A. George, and D.G. Rogers, Polym. Adv. Technol., 13, 353 (2002).

[25.] Q.H. Le , H.C. Kuan, J.B. Dai, I. Zaman, L. Luong, and J. Ma, Polymer, 51, 4867 (2010).

[26.] Y. Zhang, C. Shang, X. Yang, X. Zhao, and W. Huang, J. Mater. Sci., 47, 4415 (2012).

[27.] X. Su and X. Jing, J. Appl. Polym. Sci., 106, 737 (2007).

[28.] B. Francis, V.L. Rao, S. Jose, B.K. Catherine, R. Ramaswamy, J. Jose, and S. Thomas, J. Mater. Sci., 41, 5467 (2006).

[29.] H. Kishi. Y.B. Shi, J. Huang, and A.F. Yee, J. Mater. Sci., 32, 761 (1997).

[30.] MJ. Marks and R.V. Snelgrove, ACS Appl. Mater. Inteif., 1, 921 (2009).

[31.] Y. Martinez-Rubi, B. Ashrafi, J. Guan, C. Kingston, A. Johnston, B. Simard, V. Mirjalili, P. Hubert, L. Deng, and R.J. Young, ACS Appl. Mater. Inteif., 3, 2309 (2011). I

[32.] D.J. Hourston, J.M. Lane, and N.A. MacBeath, Polym. hit., 26, 17 (1991).

[33.] X. Kornmann, R. Thomann, R. Mulhaupt, J. Finter, and L.A. Berglund, Polym. Eng. Sci., 42, 1815 (2002).

[34.] J.P. Foreman, D. Porter, S. Behzadi, K.P. Travis, and F.R. Jones, J. Mater. Sci., 41, 6631 (2006).

[35.] R.J. Varley, J.H. Hodgkina, and G.P. Simon, Polymer, 42, 3847 (2001).

[36.] J. Zhang, Q. Guo, and B. Fox, J. Polym. Sci. Polym. Phys., 48, 417 (2010).

[37.] C.M. Sahagun and S.E. Morgan, ACS Appl. Mater. Inteif, 4, 564 (2012).

[38.] J.A. Schroeder, P.A. Madsen, and R.T. Foister, Polymer, 28, 929 (1987).

[39.] X. Sheng, J.K. Lee, and M.R. Kessler, Polymer, 50, 1264 (2009).

[40.] G. Levita, S. De Petris, A. Marchetti, and A. Lazzeri, J. Mater. Sci., 26, 2348 (1991).

[41.] L.W. Hill, Prog. Org. Coat., 31, 235 (1997).

[42.] J.D. Liu, H.J. Sue, Z.J. Thomson, F.S. Bates, M. Dettloff, G. Jacob, N. Verghese, and H. Pham, Macromolecules, 41, 7616 (2008).

[43.] W. Liu, Q. Qiu, J. Wang, Z. Huo, and H. Sun, Polymer, 49, 4399 (2008).

[44.] V.L. Zvetkov, Polymer, 42, 6687 (2001). I

[45.] H.W. Siesler, Y. Ozaki, H. Kawata, and H.M. Heise, Near-Infrared Spectroscopy, Wiley-VCH, Weinheim, Germany (2004).

[46.] L. Chiao, Macromolecules, 23, 1286 (1990).

[47.] J. Mijovic, and J. Wijaya, Polymer, 35, 2683 (1994).

[48.] M. S. Heise, and G. C. Martin, J. Appl. Polym. Sci., 39, 721, (1990).

[49.] G. Odian, Principles of Polymerization, 4th ed., Wiley, New York, Chapter 2, 39-197 (2004).

[50.] J.P. Pascault and R.J.J. Williams, J. Polym. Sci. Polym. Phys., 28, 85 (1990).

[51.] A.J. Lesser and E. Crawford. J. Appl. Polym. Sci., 66, 387 (1997). I

[52.] H. Liu, A. Uhlherr, and M.K. Bannister, Polymer, 45, 2051 (2004).

Monoj Pramanik, Eric W. Fowler, James W. Rawlins School of Polymers and High Performance Materials, The University of Southern Mississippi, 118 College Drive #5217, Hattiesburg, Mississippi 39406-0001

Correspondence to: James W. Rawlins; e-mail: james.rawlins@usm.edu Contract grant sponsor: Office of Naval Research (Award No. N0001407-1-1057).

DOI 10.1002/pen.23749

Published online in Wiley Online Library (wileyonlinelibrary.com).

TABLE 1. Physical properties of 4,4'-DDS cured epoxy thermosets.

              Tensile    Young's    [T.sub.g]
Epoxy         strength   modulus   ([degrees]C)
prepolymers    (MPa)      (GPa)

DGEBA           ~88        3.2         ~215
TGAP            ~85       ~4.15        ~265
TGDDM            --       ~5.03        ~266

               [K.sub.1c]
Epoxy           (MPa x
prepolymers   [m.sup.1/2])   References

DGEBA            ~0.51        [25, 30]
TGAP             ~0.8         [31, 35]
TGDDM            ~0.51        [26, 341

TABLE 2. Sample detail with cure schedules.

Sample   Composition (all formulations are stoichiometric)
code

S1       DGEBA-based epoxy prepolymer and 3,3'-DDS
S2       DGEBA-based epoxy prepolymer and 3,3'-DDS
S3       DGEBA-based epoxy prepolymer and 3,3'-DDS
S4       DGEBA-based epoxy prepolymer and 3,3'-DDS
S5       DGEBA-based epoxy prepolymer and 3,3'-DDS

Sample                   Cure schedule
code

S1                    90[degrees]C (18 h)
S2          90[degrees]C (18 h)+150[degrees]C (20 h)
S3                    125[degrees]C (5 h)
S4       125[degrees]C (5 h) + 200[degrees]C (1 h) (a)
S5                    I75[degrees]C (5 h)

(a) Near IR spectra were acquired for durations
up to 4 h at 200[degrees]C.

TABLE 3. Initial temperature ([T.sub.i]), peak temperature
([T.sub.p]), end temperature ([T.sub.f]), heat of exotherm ([DELTA]H)
of curing of DGEBA-based epoxy prepolymer and 3,3'-DDS obtained from
DSC analysis.

Heating rate             [T.sub.i]               [T.sub.p]
                        ([degrees]C)           ([degrees]C)

1.0[degrees]C/min   108.95 [+ or -] 2.05   156.16 [+ or -] 0.23
1.5[degrees]C/min   115.05 [+ or -] 0.64   164.50 [+ or -] 0.21
2.0[degrees]C/min   121.00 [+ or -] 0.71   171.96 [+ or -] 0.18
2.5[degrees]C/min   125.80 [+ or -] 0.85   177.77 [+ or -] 0.14

Heating rate             [T.sub.f]            [DELTA]H (J/g)
                        ([degrees]C)

1.0[degrees]C/min   220.50 [+ or -] 2.83   346.65 [+ or -] 14.35
1.5[degrees]C/min   233.66 [+ or -] 1.68   347.10 [+ or -] 19.23
2.0[degrees]C/min   239.67 [+ or -] 1.75   356.95 [+ or -] 10.67
2.5[degrees]C/min   255.12 [+ or -] 0.82   363.75 [+ or -] 21.28

Heating rate             Time (min)

1.0[degrees]C/min   109.92 [+ or -] 3.08
1.5[degrees]C/min    84.44 [+ or -] 9.14
2.0[degrees]C/min    64.85 [+ or -] 9.03
2.5[degrees]C/min    51.21 [+ or -] 0.71

TABLE 4. Epoxy conversion and glass transition temperature samples
S1-S5 as well as at the time of vitrification of each cure profiles
during cure of stoichiometric blend of DGEBA-based epoxy prepolymer
and 3,3'-DDS.

Cure condition          Time           Epoxy         [T.sub.g]
                                   conversion (%)   ([degrees]C)
                                                     (via DSC)

90[degrees]C            13 h           74.83           100.38
90[degrees]C (S1)       18 h           77.89           113.04
90[degrees]C +       18 h + 2 h        96.04           162.00
150[degrees]C
90[degrees]C +       18 h + 20 h       97.68           173.61
150[degrees]C (S2)
125[degrees]C            4 h           85.24           140.69
125[degrees]C (S3)       5 h           86.62           143.35
125[degrees]C +       5 h + 1 h        99.47           175.24
200[degrees]C (S4)
175[degrees]C            1 h           95.03           155.21
175[degrees]C (S5)       5 h           99.32           175.48

TABLE 5. Storage modulus (E'), [T.sub.g], [M.sub.c], [v.sub.c],
full width at half height (FWHH) of DGEBA-based epoxy prepolymer,
and 3,3'-DDS (stoichiometric ratio) thermosets.

Sample     Epoxy         E' (GPa) at          [T.sub.g]
         conversion      25[degrees]C        ([degrees]C)
           (NIR)                             peak of tan
                                               [delta]

S1         77.89%     2.65 [+ or -] 0.10   128 [+ or -] 1.7
S3         86.62%     2.62 [+ or -] 0.03   146 [+ or -] 0.6
S2         97.68%     2.45 [+ or -] 0.13   177 [+ or -] 0.5
S5         99.32%     2.36 [+ or -] 0.08   183 [+ or -] 0.5
S4         99.47%     2.09 [+ or -] 0.10   184 [+ or -] 1.0

Sample       E' (MPa) at           [M.sub.c]
             [T.sub.g] +            (g/mol)
            50[degrees]C

S1       10.81 [+ or -] 0.10    1154 [+ or -] 11
S3       17.44 [+ or -] 0.23     719 [+ or -] 06
S2       25.42 [+ or -] 0.59     521 [+ or -] 07
S5       26.67 [+ or -] 0.17     508 [+ or -] 02
S4       26.65 [+ or -] 0.29     511 [+ or -] 05

Sample       [v.sub.c]               FWHH
           (mol/[m.sup.3])        (tan[delta])

S1         961 [+ or -] 09     16.75 [+ or -] 0.3
S3        1491 [+ or -] 12     16.25 [+ or -] 0.4
S2        2038 [+ or -] 27     12.50 [+ or -] 0.4
S5        2113 [+ or -] 08      9.98 [+ or -] 0.4
S4        2107 [+ or -] 19      9.96 [+ or -] 0.5
COPYRIGHT 2014 Society of Plastics Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Pramanik, Monoj; Fowler, Eric W.; Rawlins, James W.
Publication:Polymer Engineering and Science
Article Type:Report
Date:Sep 1, 2014
Words:9809
Previous Article:Effect of core-shell zinc hydroxystannate nanoparticle-organic macromolecule composite flame retardant prepared by masterbatch method on...
Next Article:Effectiveness of a backward mixing screw element for glass fiber dispersion in a twin-screw extruder.
Topics:

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters